Sync main-xavier with upstream main

[leonardoxr]

Summary

• Merge upstream main from pingdotgg/t3code into main-xavier.

• Preserve fork features while resolving upstream structural changes, including assistant copy formatting, thread search, queued follow-ups, and related compatibility updates.

Validation

• bun fmt
• bun lint (warnings only)
• PATH="/Users/xavier/.cache/codex-runtimes/codex-primary-runtime/dependencies/node/bin:$PATH" bun typecheck

Note

Add queued follow-ups to threads with background reactor, search highlighting, and markdown copy format

• Introduces a queuedFollowUps feature on threads: new contracts schemas, commands, and events for enqueue/update/remove/reorder/send-failed/send-error-cleared lifecycle; decider, projector, and projection pipeline all updated to persist and project follow-up state into a new projection_thread_queued_follow_ups DB table (migration 031).
• Adds QueuedFollowUpReactor, a background service that reads the orchestration read model, dispatches thread.turn.start for eligible queued follow-ups, records pending dispatches, and emits thread.queued-follow-up.send-failed on errors.
• Adds thread search: buildThreadSearchIndex and findThreadSearchLookupState index timeline rows (messages, plans, work entries) and find matches; ThreadSearchBar provides keyboard-navigable UI; createThreadSearchHighlightRehypePlugin and renderHighlightedText highlight matches in rendered markdown and plain text.
• Adds an assistantResponseCopyFormat setting ('markdown' | 'plain-text') with a full AST-based markdownToPlainText renderer (LRU-cached) so users can copy assistant responses as plain text.
• Adds a ComposerQueuedFollowUpsPanel component with drag-and-drop reordering of queued follow-up items.
• Risk: migration 031 adds a new table with a foreign key cascade on projection_threads; existing orchestration read model consumers must handle the new queuedFollowUps array field on threads.

📊 Macroscope summarized 01a42b9. 40 files reviewed, 7 issues evaluated, 1 issue filtered, 2 comments posted

🗂️ Filtered Issues

apps/server/src/persistence/Layers/ProjectionThreadQueuedFollowUps.ts — 0 comments posted, 1 evaluated, 1 filtered

• line 142: The replaceByThreadId function at lines 139-159 performs a delete followed by multiple inserts without transaction wrapping. If the delete succeeds but any subsequent insertQueuedFollowUpRow call fails, the database will be left in an inconsistent state: the original follow-ups are deleted, but only a partial set of new follow-ups exist. This violates the atomicity semantics implied by "replace" - the operation should either fully succeed or leave the data unchanged. Consider wrapping the delete and inserts in a SQL transaction using SqlClient.withTransaction or similar. [ Failed validation ]

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: dab19cbc-a96b-4972-8dea-44b7e6154591

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

Tip

💬 Introducing Slack Agent: The best way for teams to turn conversations into code.

Slack Agent is built on CodeRabbit's deep understanding of your code, so your team can collaborate across the entire SDLC without losing context.

• Generate code and open pull requests
• Plan features and break down work
• Investigate incidents and troubleshoot customer tickets together
• Automate recurring tasks and respond to alerts with triggers
• Summarize progress and report instantly

Built for teams:

• Shared memory across your entire org—no repeating context
• Per-thread sandboxes to safely plan and execute work
• Governance built-in—scoped access, auditability, and budget controls

One agent for your entire SDLC. Right inside Slack.

👉 Get started

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[macroscopeapp]

🟢 Low lib/markdownPlainText.ts:1

The FENCED_CODE_BLOCK_REGEX requires a newline before the closing fence, so empty fenced code blocks like "```\n```" fail to match and leave the fence markers in the output. Consider using (?:[\s\S]*?\n)? to make the content line optional.

Suggested change

	const FENCED_CODE_BLOCK_REGEX = /(^|\n)(`{3,}|~{3,})[^\n]*\n([\s\S]*?)\n\2(?=\n|$)/g;
	+const FENCED_CODE_BLOCK_REGEX = /(^|\n)(`{3,}|~{3,})[^\n]*\n((?:[\s\S]*?\n)?)\2(?=\n|$)/g;

🤖 Copy this AI Prompt to have your agent fix this:

In file apps/web/src/lib/markdownPlainText.ts around line 1:

The `FENCED_CODE_BLOCK_REGEX` requires a newline before the closing fence, so empty fenced code blocks like `"[code fence]\n[code fence]"` fail to match and leave the fence markers in the output. Consider using `(?:[\s\S]*?\n)?` to make the content line optional.

Evidence trail:
apps/web/src/lib/markdownPlainText.ts line 1 (REVIEWED_COMMIT): regex `/(^|\n)(`{3,}|~{3,})[^\n]*\n([\s\S]*?)\n\2(?=\n|$)/g` — the literal `\n` between group 3 and `\2` mandates a newline before the closing fence. For input `"[code fence]\n[code fence]"`, after consuming the opening fence and the newline, the remaining text is just `` [code fence] `` with no preceding newline, so the regex cannot match.

[macroscopeapp]

🟡 Medium lib/markdownPlainText.ts:14

The EMPHASIS_REGEXES array handles **, __, ***, ___, and ~~, but omits single emphasis markers *text* and _text_ for italic. When italic text like *important* appears in the markdown, the asterisks are preserved in the output instead of being stripped, so plain text derived from search indexing contains incorrect characters.

-const EMPHASIS_REGEXES = [/(\*\*\*|___)(.*?)\1/gs, /(\*\*|__)(.*?)\1/gs, /(~~)(.*?)\1/gs] as const;

🤖 Copy this AI Prompt to have your agent fix this:

In file apps/web/src/lib/markdownPlainText.ts around line 14:

The `EMPHASIS_REGEXES` array handles `**`, `__`, `***`, `___`, and `~~`, but omits single emphasis markers `*text*` and `_text_` for italic. When italic text like `*important*` appears in the markdown, the asterisks are preserved in the output instead of being stripped, so plain text derived from search indexing contains incorrect characters.

Evidence trail:
apps/web/src/lib/markdownPlainText.ts line 14: `const EMPHASIS_REGEXES = [/(\*\*\*|___)(.*?)\1/gs, /(\*\*|__)(.*?)\1/gs, /(~~)(.*?)\1/gs] as const;` — only matches `***`, `___`, `**`, `__`, `~~` but not `*` or `_` as delimiters. Lines 31-33 show the for-loop applying these regexes. Single-emphasis italic markdown like `*text*` or `_text_` would not be matched by any regex, leaving markers in the plain text output.